Hacker News Reader

Comment by millstone | Hacker News Reader

Yes, it is not immediately obvious how mass ought to be undershood in the context of relativity. First attempts treated mass as a directed vector, with transverse and longitudinal components for an accelerated particle (now we would say that the mass is fixed, it is the momentum that changes non-linearly and just use F=dp/dt).

Mass should rightly be considered more inherent in SR that energy. Energy in relativity is coordinate dependent, but mass is not: all observers agree on a system’s mass, because it is invariant.

Re: the last paragraph, yes you can project the four vector onto time and space, and you get energy and momentum respectively. So energy is not the four vector itself but its projection onto the time axis (and mass is of course its magnitude).

View on HN

This. The quantity that “truly” exists is the energy-momentum 4-vector P. A vector exists independently of any particular reference frame. Its components, on the other hand, depend on such a reference frame.

Seen from a given reference frame, the temporal component of P is the energy E, its spatial components are the momentum 3-vector p, and its magnitude is the rest mass |P| = sqrt(E^2 - p^2) = m. The minus sign rather than plus sign is due to the Lorentzian signature. It is very helpful to visualize these quantities geometrically as 4-vectors in 4-dimensional space.

Also, the magnitude of the momentum 4-vector is a Lorentz scalar, meaning it is invariant. Hence the rest mass m = |P| is invariant.

(Relativistic) mass is totally frame dependent in SR. That's why SR normally talks about the rest or invariant mass, which is the mass in a special frame, the rest frame of the object.

You can't construct a reference frame such that a rotating object is at rest.

You can construct a reference frame such that a rotating object is rotating but not translating. You can also construct a non-inertial reference frame such that it's not even rotating.

Not quite, the laws of physics don't apply in a rotating reference frame. For example the speed of light stops being a constant.

So, if you wanted to use one you would need to redo all your math and effectively use a non rotating frame in the middle of all your calculations, or add fictitious forces and flexible constants.

> Not quite, the laws of physics don't apply in a rotating reference frame.

Yes they do and it's precisely the point of GR.

Consider a reference frame rotating at say 1 RPM. At 1 light year from the center a 'stationary' particle in a non rotation reference frame is now moving in a circle at 2 * π light years per minute. At 2 light years from the center it's moving 4 * π light years each minute and that continues the further from the center you get.

GR is fine with most reference frames as long as it's translation. Rotation is however not ok.

GR is fine with that - what will happen is that there'll be a sign change in the metric (the 'time' coordinate will become spacelike, whereas the 'angular' coordinate will become timelike), making it impossible to sit at a fixed value of the rotating frame's angular coordinate (which would correspond to faster-than-light motion in the non-rotating, inertial frame).

That's just a notation for doing exactly what I am describing.

GR deals with curved spacetime, but the only way to work in rotating spacetime is by changing how you calculate what's going on. Which means your calculations must be mapped and don't generalize.

EX: Try and do an actual calculation for say two electrons hitting each other at say 0.5c. In rotating, non rotating inertial, and non rotating non-inertial reference frames of your choice.

So, if you wanted to use one you would need to redo all your math and effectively use a non rotating frame in the middle of all your calculations, or add fictitious forces and flexible constants.

That's what the Christoffel symbols already do - it's part of the formalism, there's nothing to add!

Reply on news.ycombinator.com